[PIG-4723] Piggybank Over "last_value" UDF not working as expected.

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: piggybank
• Assignee: Chandra Prakash Bhagtani
• Reporter: Chandra Prakash Bhagtani
• Created: 01/Nov/15 06:54
• Updated: 01/Nov/15 07:02

Description

Sample input:
2014 20150801 REGISTERED
2014 20150802 NOTREGISTERED
2015 20150801 REGISTERED
2015 20150802 NOTREGISTERED

Pig Script:

A = load 'check_last_val_string.data' as (year:int,registered_date:chararray,prop:chararray);
B = group A by year;
C = FOREACH B

{ order1 = ORDER A by registered_date; generate flatten(Stitch(order1,Over(order1.prop,'last_value'))); }

;
dump C;

Output:
(2014,20150801,REGISTERED,REGISTERED)
(2014,20150802,NOTREGISTERED,NOTREGISTERED)
(2015,20150801,REGISTERED,REGISTERED)
(2015,20150802,NOTREGISTERED,NOTREGISTERED)

Expected Output:

(2014,20150801,REGISTERED,NOTREGISTERED)
(2014,20150802,NOTREGISTERED,NOTREGISTERED)
(2015,20150801,REGISTERED,NOTREGISTERED)
(2015,20150802,NOTREGISTERED,NOTREGISTERED)